Image Partitioning for Progressive Encoding Bandwidth Reduction

Solution Overview

Problem

Current progressive image encoding methods are resource-intensive, requiring significant CPU resources and network bandwidth, and do not efficiently handle low-bandwidth scenarios, leading to suboptimal image delivery in remote desktop and application delivery systems.

Innovation Solution

The method involves partitioning an image into non-overlapping vertical and horizontal rows, transmitting and interlacing sub-images to approximate the original image, reducing computational resources and bandwidth requirements, while interpolating missing pixels to achieve a complete image representation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If traditional progressive image encoding is used, then image quality is improved, but CPU resource consumption increases significantly

Engineering Contradiction:

Improveimage qualityVSAvoidCPU resource consumption

Core Design Contradiction:

Manufacturing precisionVSUse of energy by moving object

Solution Approach 1:

The image is divided into multiple sub-images or blocks that can be independently encoded and transmitted. Instead of encoding the entire image at once, the system partitions the image data into manageable segments, allowing progressive transmission and reducing the computational burden on the CPU at any given moment while maintaining image quality through systematic reconstruction of the complete image from these segments

Inventive Principle:

Principle #1Segmentation

2Manufacturing precision

If JPEG XR encoding is used, then image quality is improved, but encoding and decoding time increases

Engineering Contradiction:

Improveimage qualityVSAvoidencoding and decoding speed

Core Design Contradiction:

Manufacturing precisionVSProductivity

Solution Approach 1:

The image is divided into multiple sub-images or blocks that can be independently encoded and transmitted. Instead of encoding the entire image at once, the system partitions the image data into manageable segments, allowing progressive transmission and reducing the computational burden on the CPU at any given moment while maintaining image quality through systematic reconstruction of the complete image from these segments

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system transmits a compressed or lower-quality version of the image first, then progressively refines it by transmitting additional detail information. This allows the receiver to display an image quickly with partial information, then improve quality over time as more data arrives, rather than waiting for complete encoding of the full-resolution image

Inventive Principle:

Principle #16Partial or excessive action

3Quantity of substance

If dual-encoding technique (low quality then high quality) is used, then bandwidth efficiency is improved, but visual quality progression deteriorates

Engineering Contradiction:

Improvenetwork bandwidth usageVSAvoidvisual quality progression

Core Design Contradiction:

Quantity of substanceVSManufacturing precision

Solution Approach 1:

The image is divided into multiple sub-images or blocks that can be independently encoded and transmitted. Instead of encoding the entire image at once, the system partitions the image data into manageable segments, allowing progressive transmission and reducing the computational burden on the CPU at any given moment while maintaining image quality through systematic reconstruction of the complete image from these segments

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different regions or blocks of the image are encoded with different quality levels based on their importance. Critical areas receive higher quality encoding while less important areas use lower quality encoding, allowing efficient bandwidth utilization while maintaining acceptable visual quality progression as data is progressively received

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS8750633B2Methods and systems for approximating progressive image encoding using image partitioning

Publication Date: 2014.06.10 CITRIX SYSTEMS INC

AI summary

An image partitioner that executes on a local computer to vertically partition pixels of an original image into one or more non-overlapping vertical rows of image pixels, and horizontally partition the original image pixels into one or more non-overlapping horizontal rows of image pixels. The image partitioner can then separate the original image into a first sub-image that includes pixels of a first horizontal row and pixels of a first vertical row. The image partitioner can then separate the original image into a second sub-image that includes pixels of either the first horizontal row and a second vertical row, or a second horizontal row and the first vertical row. The first sub-image and second sub-image are then transmitted to a remote computer that interlaces the first sub-image with the second sub-image to create a first combined image which approximates the original image on the local computer.